Case study guidelines: The case study is an exercise in a) integrating the concepts learnt in class, with an in-depth focus in a single area. b) understanding the *research* and *engineering* issues in a focus area and a crisp articulation of state-of-the-art and research/engineering directions being investigated in the community. No fluff. Suggested length ~ about 10 pages. c) reading key technical research papers, IETF RFCs and drafts, participating in IETF mailing lists etc d) making a creative attempt at any open research problems to suggest solution directions/approaches. Well articulated solution approaches/breakthroughs can get anywhere from upto 5 bonus points to an outright A in the course. GUIDELINES: - The case study proposal should have: - List of sub-topics to be covered - List of papers, IETF working groups, RFCs you plan to read - The final case study writeup (of about 10 pages) should clearly specify: - the scope of the problem you are studying - detail all the fundamental design issues/subproblems involved - solution approaches to each of these issues (in isolation) - solutions adopted in practice - conclusions stating the open issues involved, and possible solution directions [ I suggest building up a template/skeleton writeup in a sections+bullets format and rework the skeleton till you get the right organization of ideas. Only after this start filling in the flesh.] EVALUATION CRITERIA: a) the depth/breadth of issues investigated b) clarity of thought and presentation c) quality of critique of the state-of-the-art and insight in presenting open issues d) Well articulated solution approaches/breakthroughs can get anywhere from upto 5 bonus points to an outright A in the course. [NOTE: Badly written reports or reports with more fluff than clear technical exposition will lose points in the evaluation.] CASE STUDY PROPOSAL: I need a proposal from you by the proposal DUE DATE (March 4th). It should contain the following: - Description of topic (the following bulletized description style is preferred). - List of papers/RFCs/Internet drafts proposed to be read. - Upto 2 students can collaborate on a case study project, provided the proposed work is substantial enough (in terms of depth/breadth} I will give feedback regarding your preproposal asap. The case studies are finally due on April 6th. Suggested Resources: ==================== - Course Resources Homepage: https://sites.ecse.rpi.edu/~koushik/shivkuma-teaching/sp2000/resources.html - IETF Active working groups: http://www.ietf.org/html.charters/wg-dir.html - Search for books on Amazon.com - Search for topics on google.com or go through the yahoo index to see what resources are available on the Internet - Researchers' home pages contain copies of papers SUGGESTED TOPICS: {you are free to suggest your own topic for inclusion here or for your case study} - Security issues and protocols for the Internet: - how is Internet security implementation different from that in other technologies, - IPSec architecture, - Key encryption/authentication techniques, - Firewalls, digital signatures/certificates - Research focus: Scalability and Manageability of the Internet security architecture - Mobility and roaming issues/protocols for the Internet - data networking protocols for wireless (MACAW, CDPD,bluetooth) - Roaming issues (PPTP draft, RFC 2194) - Mobile IP issues/solutions - Mobile TCP - Satellite-based access - 2.5G and 3G technologies for data Research focus: What are the fundamental bottlenecks to high-speed wireless data connectivity and what approaches are being taken to eradicate those bottlenecks? - Beyond best effort: - Basic concepts: admission control, scheduling, policing - Integrated services & RSVP (RFCs) vs ATM CBR/VBR - differentiated services (RFCs, IDs) vs frame-relay/ATM svcs - MPLS, traffic engineering and QoS routing in the Internet - Research focus: What strategies can be taken for provisioning, traffic engineering, deployment, service design as bandwidth becomes more and more cheaper. - Internet Pricing: - Economic basis of pricing structures - Current state of access, leased-line pricing - Internet pricing proposals: VArian/Mackie Mason's usage/congestion based pricing, Dave Clark's contracting framework, Kelly's analytical work on contract-pricing and congestion pricing,General equilibrium pricing models of Texas guys. - Interplay of pricing with Differentiated services architecture. Research focus: comparison of several schemes from a deployment perspective - Virtual private networks: - VLANs - Remote access (PPTP, L2TP) - Frame-relay and leased line (T1- or T3 line) based VPNs - CPE-based VPNs vs network-based VPNs - VPN internet drafts - Tunneling: GRE tunnels, IPSEC tunnels, IP-in-IP tunnels, L2TP tunnels, MPLS-tunnels - ATM based VPNs (MPOA vs MPLS) - Research focus: Addressing, routing and management issues - Addressing Issues in the Internet: - IPv6 vs IPv4 addressing - ATM vs IPv4 addressing - Address space management: IANA policies, CIDR protocol, subnet-masks, proxy-ARP - Address translation: NAT - Addressing for mobile hosts, virtual private networks - Security issues with addressing - Error Control, Reliability and Flow control techniques: - Bit-errors vs packet errors - Error detection/correction codes: parity, hamming codes, rectangular codes, CRC, BCH and Reed Solomon codes for error correction, Convolution coding techniques - Packet error detection techniques: Sequence numbers, three-way handshaking, incarnation numbers, acks/naks, bitmaps, selective acks, timeout, retransmission techniques. - Window-based control vs Rate-based control - Explicit rate, bit-based feedback vs implicit feedback - Window increase/decrease schemes, asymptotic analysis - Reliable Multicast and Multicast congestion control issues: ack implosion, aggregation of feedback, Probabilistic timer backoff, receiver-based control, overhead analysis, different notions of reliability, FEC-based reliability/congestion management schemes, TCP-friendliness/stability/efficiency constraints on congestion schemes, Steve McCanne's RLM scheme, Work by Bolot, Rizzo, Jon Crowcroft, UMass - Routing Issues in the Internet: - Intra-domain (RIP, OSPF) - Inter-domain routing (EGP, BGP-4, IS-IS). Critique of issues. - Explicit and QoS routing in the Internet vs PNNI (ATM) - MPLS Traffic Engineering and Constraint-based routing - Multicast inter-domain routing issues - Multicast issues/protocols in the Internet - IGMP (including current work on IGMPv3) - Intra-domain routing: dense mode (PIM-DM, MOSPF, DVMRP), Sparse mode (CBT, PIM-SM) - Inter-domain routing: IDMR, MBGP work in IETF - Reliable and Unreliable transport protocols: RTP, PGM, - Congestion control - Apps: vic, vat, wb - IP-ATM convergence/internetworking: - Fundamental internetworking problems - Multiprotocol label switching, - RFC 1577, 1483 - NHRP and IETF's ION group's work - MPOA, LANE of the ATM forum - Multicast internetworking: MARS - Reading: Anthony Alles paper (from cisco.com) - IP over WDM or SONET vs IP over ATM - Middleware for the Web - HTTP, HTTP1.1 and HTTPNG (w3c.org) - XML and other markup language technologies - What is middleware - how is it different from the layered model as a paradigm - CORBA, ORBIX, Jini architectures - Mobile Agents, Security, QoS mapping middleware - LDAP, policy servers - Web caching: work by Lixia Zhang at UCLA, UVirginia work, UC-Davis work. - Commercial offerings ? - High-Speed Router implementation issues: - Papers in IEEE Communication magazine in May (?) 1998 on this topic, Work in Washington Univ at St Louis, networking group on fast lookups, SIGCOMM 97 and 98 papers. Issues in VLSI design. - ATM switches vs IP routers - Layer 3 switching - IP switching paper in ToN, MPLS work in IETF (framework draft) - High speed switching technology: batcher-banyan networks, tera-bit switching architectures (see darpa-ito pages on current research) - Issues: High speed lookups (trie data structures), IP header processing, variable length header/packet processing, Real-time OS requirements, applicable kernel implementation issues - Commercial: Cisco IOS architecture, Juniper, Avici, Nexabit etc - Optical Networking: - Basic optical networking building blocks - WDM, DWDM - Wavelength routers - Optical MAN technology and architecture - Passive optical networks Good book: Ramaswami et al, "Optical networks" - Internet Simulation, Measurement and Modeling Issues: - Paxson's PhD thesis (or SIGCOMM papers), Bellcore papers on Self-similarity, Sugih Jamin's work on Internet Distance Maps, Floyd/Paxson paper on failure of Poisson modeling for WAN traffic and on Internet imulation issues, DIMACS (Rutgers) S-3 scalable simulations project (http://dimacs.rutgers.edu/Projects/Simulations/darpa/index.html) - Tools/traces for Internet modeling (NLANR, CAIRN, CAIDA), web performance/modeling (W3c.org/WCA) - Comment on: aspects/dimensions of Internet which require modeling - Challenge: Develop end-to-end tools for better monitoring/modeling of selected aspects of the Internet. Bonus: can earn upto 5 extra points. Fundamental work will get a straight A in the course. - Characterize in detail AT LEAST TWO OF - Windows 95, Windows 98, Windows NT, Solaris, FreeBSD's TCP implementations by reading tcpdump output. TCPdump sees protocol exchanges, but protocol processing must be inferred. - Challenge: build a tool which allows characterization of TCP implmentations in terms of throughput vs packet loss rate (for a fixed RTT) based upon tcpdump output. Bonus: A in course if you succeed. - Look up Z-iteration work of Udaya Shankar at UMD. http://www.cs.umd.edu/~shankar/Z-iteration/index.html - Required reading: Sigcomm'98 paper on TCP modeling by Padhye et al, Any relevant material from IETF's tcpimpl working group, Stevens vol2 or Comer vol2 TCP implementation chapters. - TCP performance: - Issues in reliability and congestion control - TCP Tahoe, Reno, Vegas, - Proposed enhancements: fast-retransmit/recovery, SACK, ECN - Network-based enhancements: RED, ARED, FRED, Per-flow buffer management (Lucent), TCP rate control (Shiv's paper, Packeteer) - Interaction with ATM (EPD scheme, TCP over ABR/UBR: Shiv's PhD work) - Interaction with wireless/mobile networks (Hari Balakrishnan's papers) - Interaction with asymmetric channels (Papers by TV Lakshman, Hari Balakrishnan et al) Good resource: http://www.ecse.rpi.edu/Homepages/shivkuma/research/cong-papers.html - TCP and WWW modeling: - Padhye paper in SIGCOMM'98, Sally Floyd and other people's work on "TCP-friendly" behavior - Challenge: Build a similar model for transfer time of short transfers. Build simple short packet transfers to validate your model - I will get accounts for you in sites around the Internet. Bonus: if you succeed, you get an A in the course. - Survey/critique of web modeling attempts, SPECweb benchmarks, Web characterization efforts of w3c: w3c.org/WCA Good resource: http://www.ecse.rpi.edu/Homepages/shivkuma/research/cong-papers.html